The bidirectional relationship between brain network dynamics and adolescent alcohol use

Wake Forest University Health Sciences, Winston-Salem NC

Investigators

Abstract

PROJECT SUMMARY There is a high prevalence of risky drinking in adolescents. This is concerning because alcohol use during adolescence substantially increases risk of eventually developing an alcohol use disorder (AUD). Studies have begun to show that functional neuroimaging data may be useful for predicting drinking vulnerability in adolescents. Further, studies show that drinking history is associated with altered functional brain network architecture in adolescents, particularly in three key subnetworks of the brain (the Default Mode Network [DMN], Central Executive Network [CEN], and Salience Network [SN]) which together comprise the Triple Network model. Outside of the adolescent drinking literature, methods for studying time-varying dynamic functional brain networks (DFNs) have rapidly gained popularity. DFNs depict the second-to-second changes in functional brain connectivity which have now consistently been shown to support cognitive processes. Further, DFNs are excellent tools for predicting behavioral and psychiatric outcomes, suggesting that they are more meaningful models of brain function than regional activation or static functional connectivity. However, DFNs have never been used to study vulnerability to future drinking nor the effects of drinking on the brain in an adolescent population. The proposed study will be the first to apply DFNs to studying these two important topics. I will use a distinct method for analyzing DFNs in each of two aims. In the first aim, I will use sliding window correlation-based DFNs to predict risky drinking onset in 17-year-old participants of the NCANDA study using an approach which our lab has previously used to predict success in a behavioral weight loss intervention with 95% accuracy. In the second aim, I will use Hidden Semi-Markov Models (HSMMs, a novel method for interpreting brain network dynamics) to assess the longitudinal effect of risky drinking on the brains of adolescents. My preliminary analyses using HSMMs have shown that individuals who spend more time in states with strong connectivity between the SN and CEN tend to drink less over the next year of their lives. This result provides proof of concept that 1) patterns of brain network dynamics are related to future drinking vulnerability, and 2) the specific method which will be used in the second aim of the proposed project can successfully identify patterns of brain network dynamics which are relevant to alcohol use. Together, the two aims in the proposed project will yield novel insights on the biological underpinnings of adolescent alcohol use vulnerability as well as the effect that alcohol use has on the adolescent brain. We believe that the knowledge gained from these aims will ultimately help to develop more effective prevention initiatives and provide novel explanations as to how risky drinking during adolescence can put young people on a pernicious path towards eventually developing an AUD.